Chapter 1

Introduction

Abstract This chapter first, from a historical viewpoint, shows why Petri nets are a
widely used mathematical tool to investigate supervisory control of discrete-event
systems, particularly for the deadlock analysis and control of automated manufacturing systems. The advantages and disadvantages of three major deadlock resolution
strategies in the context of resource allocation systems, which are deadlock detection and recovery, deadlock avoidance, and deadlock prevention, are analyzed. A
number of subclasses of Petri nets that can model various automated manufacturing
systems are listed. Then, it reviews the existing deadlock prevention policies in the
literature for automated manufacturing systems. The policies are qualitatively evaluated and compared briefly from computational complexity, supervisor complexity,
and behavioral permissiveness. Finally, it outlines the book.

1.1 Background
A discrete-event system (DES) is a dynamical system that evolves according to
asynchronous occurrences of discrete events. The examples of DES in the real
world include a variety of man-made systems such as flexible manufacturing systems, complex computer programs, computer networks, communication systems,
unmanned urban traffic systems, and workflow systems. A DES has a discrete set of
states that may take symbolic values rather than real numbers. State transitions in
these systems occur at asynchronous discrete instants of time in response to events,
which may also take symbolic values. Usually, the relationships between state transitions and events cannot be described by differential or difference equations.
DES is a growing area that utilizes many interesting mathematical models and
techniques. A DES is usually studied at two different levels: logical and performance levels. The models of the former are used to describe qualitative properties
and control the sequences of events in a DES. The timing of event occurrences is ignored. At this level, typical problems are the avoidance of forbidden states or event
sequences for the purpose of deadlock avoidance or liveness enforcement. The per-

1 Introduction

formance models deal with quantitative properties and aim to control the temporal
behavior of a DES. In this case, the typical problems include the satisfaction of timing constraints, scheduling, and, particularly, optimization of some key performance
criteria of a DES, for example, the production rate of a manufacturing system.
At the logical level, the most interesting and original approach to the control of
a DES is supervisory control theory (SCT). The seminal theory by P. J. Ramadge
and W. M. Wonham [6567] considers a DES as a generator of a formal language.
Its behavior can be controlled by a supervisor that prevents event occurrences in
order to satisfy a control specification. SCT aims at providing a comprehensive and
general framework that can deal with the control of DES represented by automata. It
is concerned with a qualitative treatment with a control flavor of the discrete world.
In DES literature, a system to be controlled is usually called a plant. If it is modeled with a Petri net, the resultant Petri net is called a plant (Petri) net model. It is
likely that the behavior of a plant may violate some constraints that must be enforced
to the system. As a result, a plant often needs to be controlled by an external agent
such that it behaves as one desires. A supervisor is referred to the external agent of
a system to be controlled. Consequently, the plant model and its supervisor together
are called a controlled system or controlled net if both take the form of Petri nets.
The framework proposed by Ramadge and Wonham is highly flexible with respect to the choice of models. The state space representation can be totally unstructured as in an automaton, or it can be structured as in a vector space, or it can be any
combination thereof. Due to the fact that the state space of a Petri net is structured as
a vector, Petri nets are widely used as a formalism in DES control theory. As stated
in [9], the popularity of Petri nets as a formalism for the modeling and control of
DES can be additionally attributed to the following reasons.
First, the well-established Petri net community that mainly consists of computer
scientists has developed a large family of Petri net models across many disciplines.
Different classes of Petri nets can represent different types of DES. Specifically,
place/transition nets can be used to represent the logical level of a DES. Deterministic timed event graphs, a subclass of Petri nets, are equivalent to (max,+)-linear
systems. More general timed deterministic and stochastic Petri nets can be used for
performance evaluation. High-level nets can offer a compact model for complex
systems. Hybrid nets can represent hybrid systems that involve both discrete and
continuous processes. Generalized stochastic Petri nets can model general Markovian processes, which play a key role in stochastic optimization in DES. It is shown
that the family of Petri nets developed in the literature can be used for simulation,
control, verification, performance analysis, scheduling, and optimization.
Second, Petri nets can be used in all stages starting from modeling to control
implementation. For example, Grafcet, a design paradigm of control programs for
programmable logical controllers, is usually considered as a variation of Petri nets.
Both plant and supervisor models can be represented with Petri nets. This feature
can greatly facilitate modeling, open-loop system analysis and synthesis, control
implementation, and closed-loop system analysis and evaluation.

1.2 Literature Review

Third, related computation can be made less extensive by fully utilizing the structural information of Petri nets. Also, Petri nets have a set of systematic mathematical
analysis tools employing linear matrix algebra.
Last but not least, the research results using Petri nets as a formalism to deal
with the modeling and control problems of DES over the past decades are very
fruitful. Although the decision power of a Petri net is not unlimited, a good variety
of DES control problems can be effectively and efficiently solved in a Petri net
formalism [59]. For example, Petri nets have proved to be very successful in dealing
with the forbidden-state problem, an important class of control specifications in
supervisory control of DES. The achievements made by Petri net researchers in
this area, however, in our own opinion, result partially from the supervisory control
theory initialized by Ramadge and Wonham. In fact, many ideas in Petri net domain
are borrowed from their theory, and most of research on Petri nets for DES has
strongly been influenced by their supervisory control paradigm.
The facts mentioned above indicate that Petri nets are increasingly becoming
an important and fully-fledged mathematical model to investigate the modeling and
control of DES. In a Petri net formalism, liveness is an important property of system
safety, which is equivalent to the non-blockingness in Ramadge and Wonhams supervisory control framework. Liveness implies the absence of global or local deadlock situations in a system.
A variety of theoretical results and computational algorithms have been developed in the literature to assess the liveness of certain classes of Petri nets. The liveness assessment can be performed by verifying the satisfiability of certain predicates
on siphons, a well-known structural object in Petri nets. One of the most interesting
past developments is the use of such structural objects to derive liveness-enforcing
(Petri net) supervisors for DES.
However, the power of the siphon-based liveness-enforcing approaches is degraded and deteriorated by the fact that siphons number in a Petri net grows quickly
beyond practical limits and often grows exponentially with respect to the net size.
They suffer from the computational complexity problem since it is known that in
general the complete siphon enumeration in a Petri net is NP-complete. Furthermore, they usually lead to a much more structurally complex liveness-enforcing
supervisor than the plant net model that is originally built. This book attempts to
show (1) how Petri nets can be used to deal with deadlock control problems and
(2) how the new concept of elementary siphons in a Petri net improves the existing
deadlock control policies.

1.2 Literature Review

Deadlocks have been extensively investigated in computer operating systems [2, 13,
2730, 32, 40, 60, 63, 74]. In general, they are an undesirable situation in a resource
allocation system. Their occurrence implies the stoppage of the whole or partial system operation. In a production system, for example, deadlocks and related blocking

1 Introduction

phenomena often cause unnecessary costs such as long downtime and low utilization of some critical and expensive resources, and may lead to catastrophic results in
highly automated systems, e.g., semiconductor manufacturing systems. Therefore,
it is necessary to develop an effective control policy to make sure that deadlocks
never occur in these systems. Over the last two decades, a great deal of research has
been focused on solving deadlock problems in DES, resulting in a wide variety of
approaches. This section is not intended to present a comprehensive overview of the
deadlock control approaches in the literature. We instead concentrate on the most
closely related approaches that are developed based on Petri nets.
The methods derived from a Petri net formalism for dealing with deadlocks either
preclude the possibility of deadlock occurrence by breaking some necessary conditions for a deadlock to arise or detect and resolve a deadlock when it occurs. Generally, these deadlock resolution methods are classified into three strategies: deadlock
detection and recovery [49, 85], deadlock avoidance [1, 5, 22, 34, 35, 80, 8284], and
deadlock prevention [19, 23, 24, 47, 52, 87].
A deadlock detection and recovery approach permits the occurrence of deadlocks. When a deadlock occurs, it is detected and then the system is put back
to a deadlock-free state, by simply reallocating the resources. The efficiency of
this approach depends upon the response time of the implemented algorithms
for deadlock detection and recovery. In general, these algorithms require a large
amount of data and may become complex when several types of shared resources
are considered [1].
In deadlock avoidance, at each system state an on-line control policy is used
to make a correct decision to proceed among the feasible evolutions. The main
purpose of this approach is to keep the system away from deadlock states. Aggressive methods usually lead to higher resource utilization and throughput, but
do not totally eliminate all deadlocks for some cases. In such cases if a deadlock
arises, suitable recovery strategies are still required [49, 80, 85]. Conservative
methods eliminate all unsafe states and deadlocks, and often some good states,
thereby degrading the system performance. On the other hand, they are intended
to be easy to implement.
Deadlock prevention is considered to be a well-defined problem in DES literature. It is usually achieved by using an off-line computational mechanism to
control the request for resources to ensure that deadlocks never occur. The goal
of a deadlock prevention approach is to impose constraints on a system to prevent
it from reaching deadlock states. In this case, the computation is carried out offline in a static way and once the control policy is established, the system can no
longer reach undesirable deadlock states. A major advantage of deadlock prevention algorithms is that they require no run-time cost since problems are solved in
system design and planning stages. The major criticism is that they tend to be too
conservative, thereby reducing the resource utilization and system productivity.
In the early work of Petri nets as a DES formalism, deadlock prevention is
achieved by configuring proper initial markings under which a plant Petri net model
is live. This idea can be originally traced back to the seminal works of Zhou

1.2 Literature Review

and DiCesare in the 1990s [8890]. In the last decade, a fair amount of work
in this direction has been done by Jeng, Xie, Chu, Peng, Chung, and Barkaoui
[6, 12, 4146, 94]. The liveness of a Petri net model is tied to the absence of emptiable siphons. An emptiable siphon is a set of places whose marking becomes null
during the net evolution and remains so in the subsequent markings. Most recent
work in this direction utilizes this fact to analyze and control deadlocks in a DES.
One of the distinguishing features of Ramadge and Wonhams supervisory control framework is that there is a distinct boundary between a plant to be supervised
and its supervisor such that the control implementation can be independent of the
specific technology. Unfortunately, this boundary is not clearly shown in the work
that was done in the early days of Petri nets as a DES formalism. In a deadlock
resolution domain, the situation was changed after the seminal work of Ezpeleta et
al. [19] and Lautenbach et al. [51], where liveness is enforced by adding monitors,
also called control places, to prevent siphons from being emptied. This implies that
both a plant and its supervisor are unified in a Petri net formalism. In addition, the
significance of their work lies in the fact that a plant and its supervisor are successfully separated so that control implementation technology for the latter can be
independently developed.
The success of separating a plant and its supervisor in a Petri net formalism
becomes a spur that attracts much attention. Xing et al. [87] develop a deadlock prevention policy for a class of Petri nets, which is called Production Petri Nets, where
the plant net model consists of resource places and production sequences. A deadlock structure is defined, which consists of a set of transitions. The set of resources
used in the output places of the transition set is equal to the set of resources used
in the input places of the transition set. The system is led to a deadlock state if the
number of resources used by the deadlock structure equals the capacity of the resource. A control policy is accordingly developed by adding monitors, ensuring that
for each involved resource, the deadlock structure always demands less resources
than that the system has. Furthermore, the policy is minimally restrictive, i.e., it is
optimal or maximally permissive.
As gradually recognized, the work by Ezpeleta et al. [19] suffers from a number
of problems: application coverage, behavior permissiveness, computational complexity, and structural complexity. First of all, the policy in [19] can deal with only
S3 PR, a class of Petri nets. It cannot model a manufacturing system with assembly and disassembly operations since an S3 PR is composed of state machines and
resources and a state machine cannot represent assembly and disassembly operations. Second, the policy, in a general case, cannot lead to a maximally permissive
supervisor. Third, the development of the policy depends on the complete siphon
enumeration of a plant model. Such enumeration is expensive or impossible if the
size of the plant is large since the number of siphons in a net grows exponentially
fast with respect to the net size [18, 50]. The structural complexity problem of the
supervisor results from the fact that for each strict minimal siphon in the plant net
model, a monitor has to be added to prevent it from being emptied. The years following 1995 have seen a great deal of attention focused on these problems.

1 Introduction

Many extensions to S3 PR nets have subsequently been proposed, which can be

used to model more general automated flexible manufacturing systems (FMS).
AMG (augmented marked graphs) [12]: An augmented marked graph is a Petri
net mainly composed of two sets of places: operation places and resource places.
The resultant net obtained by removing resource places and their related arcs is
a marked graph.
LS3 PR (linear system of simple sequential processes with resources) [20]: Strictly
speaking, an LS3 PR is not an extended but a restrictive version of an S3 PR. Their
difference is that a special constraint is imposed on the state machines in an
LS3 PR. A state machine in it does not contain choices at internal states that are
not the idle states. Note that idle states represent job requests.
ES3 PR (extended S3 PR) [37]: Defined by Huang et al., an ES3 PR is an ordinary
Petri net resulting from adding a set of resource places to a set of process nets
that are state machines. An S3 PMR [38], from its definition, is equivalent to an
ES3 PR in [37].
ES3 PR (extended S3 PR) [77]: Composed of a set of state machines plus a set of
resource places, this type of ES3 PR nets is more general than that defined in [37]
since it may contain arcs from transitions to resource places with their weights
perhaps being greater than one.
WS3 PSR (weighted system of simple sequential processes with several resources) [76]: It is composed of state machines and resources. The usage of
resources guarantees that they are neither destroyed nor created, i.e., conservativeness. In this sense, a WS3 PSR is a generalized Petri net.
S4 R (system of sequential systems with shared resources) [1]: An S4 R is composed of a set of state machines plus a set of resource places. Compared with
other classes of Petri nets that contain state machines, its usage of resources is
almost arbitrary and requires only conservativeness.
S4 PR [78]: An S4 PR is equivalent to an S4 R [1]. Both are developed independently.
S3 PGR2 (system of simple sequential processes with general resource requirements) [62]: An S3 PGR2 is also equivalent to an S4 R.
S PR [21]: This class of nets is a generalization of previously introduced classes
that are composed of state machines. It properly includes S4 R.
RCN (resource control nets)-merged nets [44]: An RCN-merged net includes
S3 PR and some of augmented marked graphs.
ERCN (extended resource control nets)-merged nets [86]: An ERCN-merged net
includes RCN-merged nets and some of augmented marked graphs.
ERCN -merged nets [46]: An ERCN -merged net includes ERCN-merged nets
and some of augmented marked graphs.
PNR (process nets with resources) [45]: A PNR is larger than the class of S3 PR,
augmented marked graphs, and some of ERCN-merged nets.
G-tasks [6]: A G-task is composed of acyclic state machines and a set of resource
places. The resources can be arbitrarily used as long as their conservativeness is
preserved.

1.2 Literature Review

G-systems [94]: A G-system is the most general one among all the mentioned
classes. It can properly contain each of the above classes. A G-system can model
assembly (synchronization) and disassembly (splitting) operations in an FMS.
These classes can model various resource allocation systems. Their deadlock
control policies are developed according to the relationship between liveness and
siphons.
As known, the limited behavior permissiveness is a flaw in the notable deadlock
prevention policy in [19]. Huang et al. claim that the deadlock prevention policy
developed in [36] for S3 PR is in general more permissive than the one in [19]. This
statement is not formally proved. Actually, in the opinion of the authors of this book,
it may not be possible to develop a formal proof. The statistical investigation does
support such a claim [36].
Huangs policy consists of two stages and performs the synthesis of a supervisor
in an iterative way. The first stage, called siphon control, adds monitors to the plant
model such that all the siphons in the plant are controlled. The siphon control stage is
optimal or maximally permissive in the sense that no good states are removed due to
the addition of monitors. In fact, the control of a siphon in this stage is implemented
by enforcing a generalized mutual exclusion constraint (GMEC).
The second stage aims at making the newly generated siphons controlled, which
result from the addition of the monitors in the first stage. To accelerate the convergence rate, the output arcs of the monitors added in the second stage point to only
the source transitions of the plant model.
Sometimes termed optimality, maximal permissiveness is also an important parameter of a supervisor. In Ramadge and Wonhams approach, the existence and
synthesis of an optimal non-blocking supervisor for a DES has been well addressed
in a finite automaton and formal language paradigm. The existence of a synthesis
approach for an optimal liveness-enforcing supervisor remains open until the work
in [3, 26, 79]. By using the theory of regions [3] that can derive pure Petri nets from
an automaton-based model, Uzam [79] develops an optimal liveness-enforcing supervisor synthesis method on the condition that such a supervisor exists. However, it
is difficult to understand and use. Later, by using plain and popular linear algebraic
notions, Ghaffari et al. [26] explore the conditions on the existence of an optimal
supervisor that is maximally permissive, and develop a methodology to synthesize
it.
These explicit approaches that need to generate the reachability graph of a
Petri net require memory and time at least proportional to the number of reachable
markings. Thus they are applicable to fairly small systems only. That is to say, a
plant net model has to be small-sized. Also, its initial marking must be so small that
its reachability set is limited to the computers memory and processing capability.
As a result, the computational efficiency is the Achilles heel of methods of this
kind since the complete state enumeration is needed. This is not surprising since
the theory of regions is a method to derive Petri nets from an existing automaton
model. The work in [57] develops an optimal net supervisor design method that is
based on the theory of regions. Its efficiency is improved by reducing the number of
inequality systems that are used to separate events from some unsafe states.

1 Introduction

Computational complexity has been a major problem when a deadlock prevention policy is developed [2, 61]. For a class of Petri nets, S3 PGR2 , Park and Reveliotis [62] propose a deadlock prevention policy that, originally developed under
a finite-state automaton paradigm, is polynomial. Additional deadlock avoidance
policies that are of polynomial-time complexity are presented in [21, 34]. They are
not optimal in general.
Due to the inherent characteristics of Petri nets, the development of a polynomialtime algorithm to design a liveness-enforcing monitor-based supervisor is by no
means an easy task. An efficient way of improving the computational efficiency
of a siphon-based deadlock prevention policy is the introduction of the MIP-based
deadlock detection method pioneered by Chu and Xie [12]. It was first used by
Huang et al. in [36] to design a liveness-enforcing supervisor such that the complete
siphon enumeration is successfully avoided. In this sense, this deadlock prevention
policy enjoys high computational efficiency compared with the existing ones in the
literature at that time. The MIP-based deadlock detection method is then used in [54]
and [56].
A liveness-enforcing monitor-based supervisor derived from siphons reaches its
high structural complexity when the number of siphons is large. This problem, having been recognized for a long time, has remained open for many years. By fully
utilizing the structural information in a Petri net, the work by Li and Zhou proposes
the concepts of elementary and dependent siphons in a Petri net [53, 55]. Siphons
in a Petri net can be divided into elementary and dependent ones. The latter can
be further distinguished by strongly and weakly dependent siphons with respect to
elementary ones. It is shown that the number of the elementary siphons in a net
is bounded by the smaller of place and transition counts. Moreover, a dependent
siphon can be controlled by properly supervising the number of tokens that can stay
at its elementary siphons.
The results concerning elementary siphons mentioned above can be naturally
applied to most of the siphon-based deadlock prevention policies in the literature.
For example, monitors can be added for elementary siphons only. The controllability
of a dependent siphon can be ensured by properly supervising the initial number of
tokens in the monitors that are added to its elementary siphons. That is to say, it
is possible that we do not need to explicitly add a monitor for a dependent siphon
any more. This is fully shown in [53] by an FMS example. In theory, the size of
a supervisor that is computed by using elementary siphons is as a result less than
that of the plant. Note that the method in [53] does not lower the computational
complexity and improve the behavior permissiveness compared with the policy in
[19]. On the positive side, it does lower the structural complexity of the supervisor
notably.
It is worth noting that there is an established tool inside Petri net theory, which
can be used to remove redundant monitors from a liveness-enforcing supervisor.
It is called implicit places [14, 25, 68, 73]. Implicit places have the property that
their addition to or removal from a net system does not change its behavior, i.e., an
implicit place represents redundancy. In fact, the concept of implicit places has been
proposed for many years before the existence of the structural complexity problem

1.3 Outline of the Book

of a liveness-enforcing monitor-based supervisor. Unfortunately, no work in this

direction is found in the literature except for [58].
For a dozen of years, we have witnessed that the results are much enriched in the
area of liveness-enforcing supervisory control that is based on a Petri net formalism.
On the other hand, many interesting problems remain unsolved, particularly the
four above-mentioned hurdles, i.e., application scope, behavioral permissiveness,
computational efficiency, and supervisors structural complexity. This monograph
represents the important research results that can be used to overcome these hurdles.

1.3 Outline of the Book

This monograph is intended to present a Petri net approach to deadlock resolution
of automated manufacturing systems. It focuses on the role of elementary siphons
of Petri nets in the development of a supervisor subject to liveness and other control
requirements. It is outlined as follows.
Chapter 2 introduces the basics of Petri nets as well as the necessary notations
used throughout this book. It also includes a brief comparison between Petri nets
and automata.
Chapter 3 first defines the concepts of elementary and dependent siphons in Petri
nets. Then, important results on elementary siphons such as their number in a net and
the controllability of a dependent siphon are presented. The material in this chapter
facilitates understanding of the development of deadlock prevention policies that are
based on elementary siphons. Simple examples are given to illustrate these results.
Chapter 4 first presents a novel monitor implementation of a set of generalized
mutual exclusion constraints that are divided into elementary and dependent ones, as
motivated by the concept of elementary siphons. Conditions are then derived under
which a dependent constraint is implicitly enforced. The constraint enforcement
method is applied to a deadlock prevention policy developed in [62].
Chapter 5 introduces a well-established deadlock prevention policy via typical
examples in the literature, and then shows the application of elementary siphons
to the design of structurally simple liveness-enforcing monitor-based supervisors.
The significance of elementary siphons is fully demonstrated. A few novel deadlock
control strategies are accordingly presented.
For a class of Petri nets, Chap. 6 explores the existence and synthesis method of
a liveness-enforcing monitor-based supervisor such that the controlled (net) system
is maximally permissive on the assumption that all transitions are controllable and
observable.
Chapter 7 presents and compares the existing deadlock prevention policies for
flexible manufacturing systems via a case study. The comparison is conducted from
the following points of view: computational complexity, structural complexity, and
the behavior permissiveness.

10

1 Introduction

Chapter 8 concludes this book by presenting and discussing a number of open

and interesting problems in the field of DES control using a Petri net formalism and
their relations with other DES formalisms.

1.4 Bibliographical Remarks

Before 1990, the work that used Petri nets as a formalism to deal with deadlock
problems in DES was owing to E. Roszkowska [4, 70, 71]. However, Petri nets received more and more attention from academia and industry only after the publication of the research in [5, 80, 88].
There are several survey papers and books that investigate the supervisory control
problems of DES using Petri nets: [31, 39, 59, 69]. The paper [23] is a tutorial that
surveys the deadlock control approaches in the literature. The edited volume [93]
is the first comprehensive book that is dedicated to deadlock resolution methods in
various computer-integrated systems. Other significant books published in the area
of Petri nets and manufacturing automation include [8,1517,64,81,9092]. For the
general problems of DES, the reader is referred to [7, 10, 11, 33, 48, 72, 75].

Problems
1.1. Some supervisory control problems are investigated and well addressed in the
RamadgeWonham framework but this is not the case in a Petri net domain, e.g.,
the problems involving controllability and observability of events and decentralized control. Analyze and discuss the reasons from the development history of DES
modeling and control theory. Reader can refer to [9].

References
1. Abdallah, I.B., ElMaraghy, H.A. (1998) Deadlock prevention and avoidance in FMS: A Petri
net based approach. International Journal of Advanced Manufacturing Technology, vol.14,
no.10, pp.704715.
2. Araki, T., Sugiyama, Y., Kasami, T., Okui, J. (1977) Complexity of the deadlock avoidance
problems. In Proc. 2nd IBM Symposium on the Mathematical Foundations of Computer Science, pp.229252.
3. Badouel, E., Darondeau, P. (1998) Theory of regions. Lectures on Petri Nets I: Basic Models,
Lecture Notes in Computer Science, vol.1491, W. Reisig and G. Rozenberg (Eds.), pp.529
586.
4. Banaszak, Z., Roszkowska, E. (1988) Deadlock avoidance in pipeline concurrent processes.
Podstawy Sterowania (Foundations of Control), vol.18, no.1, pp.317.
5. Banaszak, Z., Krogh, B.H. (1990) Deadlock avoidance in flexible manufacturing systems
with concurrently competing process flows. IEEE Transactions on Robotics and Automation,
vol.6, no.6, pp.724734.

References

11

6. Barkaoui, K., Chaoui, A., Zouari, B. (1997) Supervisory control of discrete event systems
based on structure theory of Petri nets. In Proc. IEEE Int. Conf. on Systems, Man, and Cybernetics, pp.37503755.
7. Ben-Naoum, L., Boel, R., Bongaerts, L., De Schutter, B., Peng, Y., Valckenaers, P., Vandewalle, J., Wertz, V. (1995) Methodologies for discrete event dynamic systems: A survey.
Journal A, vol.36, no.4, pp.314.
8. Bogdan, S., Lewis, F.L., Kovacic, Z., Mireles, J. (2006) Manufacturing Systems Control Design. London: Springer.
9. Cao, X.R., Cohen, G., Giua, A., Wonham, W.M., Van Schuppen, J.H. (2002) Unity in diversity, diversity in unity: Retrospective and prospective views on control of discrete event
systems. Journal of Discrete Event Dynamic Systems: Theory and Applications, vol.12, no.3,
pp.253264.
10. Cassandras, C.G., Lafortune, S. (1999) Introduction to Discrete Event Systems. Boston, MA:
Kluwer.
11. Cassandras, C.G., Lafortune, S. (2008) Introduction to Discrete Event Systems. Springer.
12. Chu, F., Xie, X.L. (1997) Deadlock analysis of Petri nets using siphons and mathematical
programming. IEEE Transactions on Robotics and Automation, vol.13, no.6, pp.793804.
13. Coffman, E.G., Elphick, M.J., Shoshani, A. (1971) System deadlocks. ACM Computing Surveys, vol.3, no.2, pp.6778.
14. Colom, J.M., Silva, M. (1989) Improving the linearly based characterization of P/T nets. In
Proc. 10th Int. Conf. on Applications and Theory of Petri Nets, G. Rozenberg (Ed.), Lecture
Notes in Computer Science, vol.483, pp.113145.
15. David R., Alla, H. (1992) Petri Nets and Grafcet. Englewood Cliffs, NJ: Prentice-Hall.
16. Desrocher, A.A., AI-Jaar, R.Y. (1995) Applications of Petri Nets in Manufacturing Systems:
Modeling, Control, and Performance Analysis. Piscataway, NJ: IEEE Press.
17. DiCesare, F., Harhalakis, G., Porth, J.M., Vernadat, F.B. (1993) Practice of Petri Nets in
Manufacturing. Chapman and Hall.
18. Ezpeleta, J., Couvreur, J.M., Silva, M. (1993) A new technique for finding a generating family
of siphons, traps, and st-components: Application to colored Petri nets. In Advances in Petri
Nets, Lecture Notes in Computer Science, vol.674, G. Rozenberg (Ed.), pp.126147.
19. Ezpeleta, J., Colom, J.M., Martinez, J. (1995) A Petri net based deadlock prevention policy
for flexible manufacturing systems. IEEE Transactions on Robotics and Automation, vol.11,
no.2, pp.173184.
20. Ezpeleta, J., Garca-Valles, F., Colom, J.M. (1998) A class of well structured Petri nets for
flexible manufacturing systems. In Proc. 19th Int. Conf. on Applications and Theory of Petri
Nets, Lecture Notes in Computer Science, vol.1420, J. Desel and M. Silva (Eds.), pp.6483.
21. Ezpeleta, J., Tricas, F., Garca-Valles, F., Colom, J.M. (2002) A bankers solution for deadlock avoidance in FMS with flexible routing and multiresource States. IEEE Transactions on
Robotics and Automaton, vol.18. no.4, pp.621625.
22. Ezpeleta, J., Recalde, L. (2004) A deadlock avoidance approach for non-sequential resource
allocation systems. IEEE Transactions on Systems, Man, and Cybernetics, Part A, vol.34,
no.1, pp.93101.
23. Fanti, M.P., Zhou, M.C. (2004) Deadlock control methods in automated manufacturing systems. IEEE Transactions on Systems, Man, and Cybernetics, Part A, vol.34, no.1, pp.522.
24. Fanti, M.P., Zhou, M.C. (2005) Deadlock control methods in automated manufacturing systems. In Deadlock Resolution in Computer-Integrated Systems, New York: Marcel Dekker,
pp.122.
25. Garca-Valles, F., Colom, J.M. (1999) Implicit places in net systems. In Proc. 8th Int. Workshop on Petri Nets and Performance Models, pp.104113.
26. Ghaffari, A., Rezg, N., Xie, X.L. (2003) Design of a live and maximally permissive Petri
net controller using the theory of regions. IEEE Transactions on Robotics and Automation,
vol.19, no.1, pp.137142.
27. Gligor, V., Shattuck, S. (1980) On deadlock detection in distributed systems. IEEE Transactions on Software Engineering, vol.6, no.5, pp.435440.

12

1 Introduction

28. Gold, E.M. (1978) Deadlock predication: Easy and difficult cases. SIAM Journal of Computing, vol.7, no.3, pp.320336.
29. Haberman, A. (1969) Prevention of system deadlocks. Communications of the ACM, vol.12,
no.7, pp.373377.
30. Hack, M.H.T. (1972) Analysis of Production Schemata by Petri Nets. Master Thesis, Massachusetts Institute of Technology, Cambridge, Massachusetts, USA.
31. Holloway, L.E., Krogh, B.H., Giua, A. (1997) A survey of Petri net methods for controlled
discrete event systems. Discrete Event Dynamic Systems: Theory and Applications, vol.7,
no.2, pp.151190.
32. Holt, R. (1972) Some deadlock properties of computer systems. ACM Computing Surveys,
vol.4, no.3, pp.179196.
33. Hruz, B., Zhou, M.C (2007) Modeling and Control of Discrete-Event Dynamic Systems: With
Petri Nets and Other Tools. London: Springer.
34. Hsieh, F.S., Chang, S.C. (1994) Dispatching-driven deadlock avoidance controller synthesis
for flexible manufacturing systems. IEEE Transactions on Robotics and Automation, vol.10,
no.2, pp.196209.
35. Hsieh, F.S. (2004) Fault-tolerant deadlock avoidance algorithm for assembly processes. IEEE
Transactions Systems, Man, and Cybernetics, Part A, vol.34, no.1, pp.6579.
36. Huang, Y.S., Jeng, M.D., Xie, X.L., Chung, S.L. (2001) Deadlock prevention policy based on
Petri nets and siphons. International Journal of Production Research, vol.39, no.2, pp.283
305.
37. Huang, Y.S., Jeng, M.D., Xie, X.L., Chung, S.L. (2001) A deadlock prevention policy for
flexible manufacturing systems using siphons. In Proc. IEEE Int. Conf. on Robotics and Automation, pp.541546.
38. Huang, Y.S., Jeng, M.D., Xie, X.L., Chung, D.H. (2006) Siphon-based deadlock prevention
policy for flexible manufacturing systems. IEEE Transactions on Systems, Man, and Cybernetics, Part A, vol.36, no.6, pp.21522160.
39. Iordache, M.V. (2003) Methods for the Supervisory Control of Concurrent Systems Based on
Petri Net Abstractions. Doctoral Dissertation, University of Notre Dame.
40. Isloor, S.S., Marsland, T.A. (1980) The deadlock problem: An overview. Computer, vol.13,
no.9, pp.5877.
41. Jeng, M.D., DiCesare, F. (1993) A review of synthesis techniques for Petri nets with applications to automated manufacturing systems. IEEE Transactions on Systems, Man, and
Cybernetics, Part A, vol.23, no.1, pp.301312.
42. Jeng, M.D., DiCesare, F. (1995) Synthesis using resource control nets for modeling sharedresource systems. IEEE Transactions on Robotics and Automation, vol.11, no.3, pp.317327.
43. Jeng, M.D. (1997) A Petri net synthesis theory for modeling flexible manufacturing systems.
IEEE Transactions on Systems, Man and Cybernetics, Part B, vol.27, no.2, pp.169183.
44. Jeng, M.D., Xie, X.L. (1999) Analysis of modularly composed nets by siphons. IEEE Transactions on Systems, Man, and Cybernetics, Part A, vol.29, no.4, pp.399406.
45. Jeng, M.D., Xie, X.L., Peng, M.Y. (2002) Process nets with resources for manufacturing
modeling and their analysis. IEEE Transactions on Robotics and Automation, vol.18, no.6,
pp.875889.
46. Jeng, M.D., Xie, X.L., Chung, S.L. (2004) ERCN* merged nets for modeling degraded behavior and parallel processes in semiconductor manufacturing systems. IEEE Transactions
on Systems, Man, and Cybernetics, Part A, vol.34, no.1, pp.102112.
47. Jeng, M.D., Xie, X.L. (2005) Deadlock detection and prevention of automated manufacturing systems using Petri nets and siphons. In Deadlock Resolution in Computer-Integrated
Systems, M. C. Zhou and M. P. Fanti (Eds.), pp.233-281, New York: Marcel Dekker.
48. Kumar, R. Garg, V. (1995) Modeling and Control of Logical Discrete Event Systems. Boston,
MA: Kluwer.
49. Kumaran, T.K., Chang, W., Cho, H., Wysk, R.A. (1994) A structured approach to deadlock
detection, avoidance and resolution in flexible manufacturing systems. International Journal
of Production Research, vol.32, no.10, pp.23612379.

References

13

50. Lautenbach, K. (1987) Linear algebraic calculation of deadlocks and traps. In Concurrency
and Nets, K. Voss, H. J. Genrich and G. Rozenberg (Eds.), pp.315336.
51. Lautenbach, K., Ridder, H. (1993) Liveness in bounded Petri nets which are covered by Tinvariants. In Proc. 13th Int. Conf. on Applications and Theory of Petri Nets, Lecture Notes in
Computer Science, vol.815, R. Valette (Ed.), pp.358375.
52. Lautenbach, K., Ridder, H. (1996) The linear algebra of deadlock avoidancea Petri net
approach. No.25-1996, Technical Report, Institute of Software Technology, University of
Koblenz-Landau, Koblenz, Germany.
53. Li, Z.W., Zhou, M.C. (2004) Elementary siphons of Petri nets and their application to deadlock prevention in flexible manufacturing systems. IEEE Transactions on Systems, Man, and
Cybernetics, Part A, vol.34, no.1, pp.3851.
54. Li, Z.W., Zhou, M.C. (2006) Two-stage method for synthesizing liveness-enforcing supervisors for flexible manufacturing systems using Petri nets. IEEE Transactions on Industrial
Informatics, vol.2, no.4, pp.313325.
55. Li, Z.W., Zhou, M.C. (2006) Clarifications on the definitions of elementary siphons of Petri
nets. IEEE Transactions on Systems, Man, and Cybernetics, Part A, vol.36, no.6, pp.1227
1229.
56. Li, Z.W., Hu, H.S., Wang, A.R. (2007) Design of liveness-enforcing supervisors for flexible
manufacturing systems using Petri nets. IEEE Transactions on Systems, Man, and Cybernetics, Part C, vol.37, no.4, pp.517526.
57. Li, Z.W., Zhou, M.., Jeng, M.D. (2008) A maximally permissive deadlock prevention policy
for FMS based on Petri net siphon control and the theory of regions. IEEE Transactions on
Automation Science and Engineering, vol.5, no.1, pp.182188.
58. Li, Z.W. (2009) On systematic methods to remove redundant monitors from livenessenforcing net supervisors. To appear in Computer and Industrial Engineering.
59. Moody, J.O., Antsaklis, P.J. (1998) Supervisory Control of Discrete Event Systems Using Petri
Nets. Boston, MA: Kluwer.
60. Newton, G. (1979) Deadlock prevention, detection, and resolution: An annotated bibliography. ACM SIGOPS Operating Systems Review, vol.13, no.2, pp.3344.
61. Pablo, J., Colom, J.M. (2006) Resource allocation systems: Some complexity results on the
S4 PR class. In Proc. IFIP International Federation for Information Processing, Lecture Notes
in Computer Science, vol.4229, E. Najm et al. (Eds.), pp.323338.
62. Park, J., Reveliotis, S.A. (2001) Deadlock avoidance in sequential resource allocation systems
with multiple resource acquisitions and flexible routings. IEEE Transactions on Automatic
Control, vol.46, no.10, pp.15721583.
63. Peterson, J.L., Silberschatz, A. (1985) Operating System Concepts. Reading, MA: AddisonWesley.
64. Porth, J.M., Xie, X.L. (1996) Petri Nets, A Tool for Design and Management of Manufacturing Systems. New York: John Wiley & Sons.
65. Ramadge, P., Wonham, W.M. (1987) Supervisory control of a class of discrete event processes. SIAM Journal on Control and Optimization, vol.25. no.1, pp.206230.
66. Ramadge, P., Wonham, W.M. (1987) Modular feedback logic for discrete event systems.
SIAM Journal on Control and Optimization, vol.25, no.5, pp.12021218.
67. Ramadge, P., Wonham, W.M. (1989) The control of discrete event systems. Proceedings of
the IEEE, vol.77, no.1, pp.8189.
68. Recalde, L., Teruel, E., Silva, M., (1997) Improving the decision power of rank theorems. In
Proc. IEEE Int. Conf. on Systems, Man, and Cybernetics, pp.37683773.
69. Reveliotis, S.A. (2005) Real-time Management of Resource Allocation Systems: A Discrete
Event Systems Approach. New York: Springer.
70. Roszkowska, E., Wojcik, R. (1993) Problems of process flow feasibility in FAS. In CIM in
Process and Manufacturing Industries, Oxford: Pergamon Press, pp.115120.
71. Roszkowska, E., Jentink, J. (1993) Minimal restrictive deadlock avoidance in FMSs. In Proc.
European Control Conf., J. W. Nieuwenhuis, C. Pragman, and H. L. Trentelman, (Eds.), vol.2,
pp. 530534.

14

1 Introduction

72. Silva, M., Teruel, E. (1996) A systems theory perspective of discrete event dynamic systems: The Petri net paradigm. In Symposium on Discrete Events and Manufacturing Systems,
IMACS Multiconference, P. Borne, J. C. Gentina, E. Craye, and S. El Khattabi, (Eds.), Lille,
France, pp.112.
73. Silva, M., Teruel, E., Colom, J.M. (1998) Linear algebraic and linear programming techniques
for the analysis of place/transition net systems. In Lectures on Petri Nets I: Basic Models,
Lectures Notes in Computer Science, vol.1491, W. Reisig and G. Rozenberg (Eds.), pp.309
373.
74. Singhal, M. (1989) Deadlock detection in distributed systems. IEEE Computer, vol.22, no.11,
pp.3748.
75. Thistle, J.G. (1996) Supervisory control of discrete event systems. Mathematical and Computer and Modeling, vol.23, no.1112, pp.2553.
76. Tricas, F., Martinez, J. (1995) An extension of the liveness theory for concurrent sequential
processes competing for shared resources. In Proc. IEEE Int. Conf. on Systems, Man, and
Cybernetics, pp.30353040.
77. Tricas, F., Garca-Valles, F., Colom, J.M., Ezpeleta, J. (1998) A structural approach to the
problem of deadlock prevention in processes with shared resources. In Proc. 4th Workshop
on Discrete Event Systems, pp.273278.
78. Tricas, F., Garca-Valles, F., Colom, J.M., Ezpeleta, J. (2000) An iterative method for deadlock prevention in FMS. In Proc. 5th Workshop on Discrete Event Systems, R. Boel and
G.Stremersch (Eds.), pp.139148.
79. Uzam, M. (2002) An optimal deadlock prevention policy for flexible manufacturing systems
using Petri net models with resources and the theory of regions. International Journal of
Advanced Manufacturing Technology, vol.19, no.3, pp.192208.
80. Viswanadham, N., Narahari, Y., Johnson, T. (1990) Deadlock prevention and deadlock
avoidance in flexible manufacturing systems using Petri net models. IEEE Transactions on
Robotics and Automation, vol.6, no.6, pp.713723.
81. Viswanadham, N., Narahari, Y. (1992) Performance Modelling of Automated Manufacturing
Systems. Englewood Cliffs, NJ: Prentice Hall.
82. Wu, N. Q. (1999) Necessary and sufficient conditions for deadlock-free operation in flexible
manufacturing systems using a colored Petri net model. IEEE Transactions on Systems, Man,
and Cybernetics, Part C, vol.29, no.2, pp.192204.
83. Wu, N.Q., Zhou, M.C. (2001) Avoiding deadlock and reducing starvation and blocking in
automated manufacturing systems. IEEE Transactions on Robotics and Automation, vol.17,
no.5, pp.658669.
84. Wu, N.Q., Zhou, M.C. (2005) Modeling and deadlock avoidance of automated manufacturing
systems with multiple automated guided vehicles. IEEE Transactions on Systems, Man, and
Cybernetics, Part B, vol.35, no.6, pp.11931202.
85. Wysk, R.A., Yang, N.S., Joshi, S. (1994) Resolution of deadlocks in flexible manufacturing
systems: avoidance and recovery approaches. Journal of Manufacturing Systems, vol.13, no.2,
pp.128138.
86. Xie, X.L., Jeng, M.D. (1999) ERCN-merged nets and their analysis using siphons. IEEE
Transactions on Robotics and Automation, vol.15, no.4, pp.692703.
87. Xing, K.Y., Hu, B.S., Chen, H.X. (1996) Deadlock avoidance policy for Petri-net modelling
of flexible manufacturing systems with shared resources. IEEE Transactions on Automatic
Control, vol.41, no.2, pp.289295.
88. Zhou, M.C., DiCesare, F. (1991) Parallel and sequential exclusions for Petri net modeling for
manufacturing systems. IEEE Transactions on Robotics and Automation, vol.7, no.4, pp.515
527.
89. Zhou, M.C., DiCesare, F. (1992) A hybrid methodology for synthesis of Petri nets for manufacturing systems. IEEE Transactions on Robotics and Automation, vol.8, no.3, pp.350361.
90. Zhou, M.C., DiCesare, F. (1993) Petri Net Synthesis for Discrete Event Control of Manufacturing Systems. Boston, MA: Kluwer.
91. Zhou, M.C. (Ed.) (1995) Petri Nets in Flexible and Agile Automation. Norwell, MA: Kluwer.

References

15

92. Zhou, M.C., Venkatesh, K. (1998) Modelling, Simulation and Control of Flexible Manufacturing Systems: A Petri Net Approach. Singapore: World Scientific.
93. Zhou, M.C., Fanti, M.P. (Eds.) (2005) Deadlock Resolution in Computer-Integrated Systems.
New York: MarcelDekker.
94. Zouari, B., Barkaoui, K. (2003) Parameterized supervisor synthesis for a modular class of
discrete event systems. In Proc. IEEE Int. Conf. on Systems, Man, and Cybernetics, pp.1874
1879.